Specific docking of the three T cell receptor complementary-determining regions (CDRs) onto the peptide-MHC complex is the basis of immune recognition. However, the contribution of each CDR to peptide-MHC binding is not routine.
T cell receptors (TCRs) dock onto peptide–major histocompatibility complex (peptide-MHC) ligands in a diagonal orientation, allowing the third complementarity-determining regions (CDR3s) to focus on the central peptide and the CDR1 and CDR2 loops to engage the MHC1,2. This would seem to make for a straightforward story on the mechanisms of peptide specificity and MHC restriction, respectively. In this issue of Nature Immunology, Borg et al. show that things are not quite so simple, at least for some TCRs3. Their study analyzes the binding energetics of alanine mutants in a well studied immunodominant TCR, LC13, that is specific for a peptide (FLR) from Epstein-Barr virus nuclear antigen EBNA3 bound to HLA-B8. In this system, CDR3 loops have the main role in peptide-MHC binding, whereas CDR1 and CDR2 seem to have only supporting roles.
Based on both structural and theoretical grounds, it seemed reasonable that germline-encoded CDR1 and CDR2 would in fact be involved in MHC recognition and that highly diverse CDR3s would recognize the also very diverse set of antigens. CDR2 loops are generally found over the MHC helices, consistent with early observations that variable α (Vα) regions are important in the distinction of MHC class I versus class II products4. Various mutagenesis and binding studies even led to a proposed two-step binding model, in which CDR1 and CDR2 loops engage MHC helices, followed by CDR3 reorganization and binding of peptide5. However, based on structures of BM3.3 TCR–peptide-Kb complexes, it has been argued that a strict order of CDR binding for all TCRs would not be likely6, an argument supported by the LC13 story.
The findings of Borg et al. now place more emphasis on a critical, but perhaps underappreciated, function of CDR1 and CDR2: to shape the conformation of the CDR3 loops into a peptide-MHC-competent binding state. The study was based on alanine-scanning mutagenesis to demonstrate the presence of binding energy 'hotspots'. Alanine scans are useful in determining the binding free energy associated with a particular side chain. An excellent review7 describes important caveats of the approach, including possible complicating effects of mutations on the conformation or dynamics of a protein. Another problem with alanine scans is that mainchain interactions are not directly accessed and the corresponding binding free energy contributed by some regions may therefore not be included in the overall energy landscape. Nevertheless, it is possible with appropriate structural information to propose potential mechanistic effects of mutations. If there were ever a real biological reason to determine the relative energetic contributions of different regions of a protein, it seems that TCR–peptide-MHC interactions qualify. The processes that involve positive and negative selection, antigen responsiveness, alloreactivity and autoimmune reactions all hinge in a quantitative way on these interactions.
Two TCRs have now been subjected to alanine scanning: the 2C TCR with ligands QL9-Ld (ref. 8) and SIYR-Kb (ref. 9), and the LC13 TCR with the FLR–HLA-B8 ligand3. In the case of the 2C TCR, the salient finding was that CDR1 and CDR2 loops contribute most of the binding free energy. This finding differs from the situation with LC13, in which CDR3 loops contribute most of the binding free energy (Fig. 1). What are the main differences between these TCRs that might account for the results? Perhaps the most obvious difference is the length of their CDR3s: LC13 is four residues longer in CDR3α and two residues longer in CDR3β. Thus, LC13 loops span a larger area, and individual residues are important not only in direct contact with the peptide and/or MHC but also in configuring the loops for such contacts. In contrast, the 2C CDR3 loops are somewhat uninteresting; seven residues are glycines or alanines and the CDR3s are largely devoid of residues that have been found to be prevalent in hotspots (Trp, Arg, Tyr, Asp, Pro and His)7. The 2C TCR, however, does contain nine of these residues in CDR1 or CDR2 loops, and five of these were identified as hotspots. Furthermore, in the 2C TCR–SIYR-Kb structure, many contacts are between mainchain atoms of CDR3s (especially CDR3β, in which there are four glycines in a row) and peptide-MHC10.
In LC13, CDR1α and CDR1β residues seem to be more important in stabilizing or 'supporting' the CDR3 loops (Fig. 1a). Of CDR2 residues, only Leu50α had a substantial effect on binding, due possibly to stabilization of CDR1α or to contact with MHC. As the authors point out, this single interaction seems hardly enough to account for MHC docking or restriction. Two residues (His33α and His48α) that flank CDR1α and CDR2α were important, not in their interaction with MHC but with the tyrosine at peptide residue 7 (P7-Tyr). Notably, these two residues are unique to the Vα26 region, in part explaining the dominance of this Vα in EBV-specific T cells.
The distinct findings regarding the CDR location of energetic hotspots raise the issues of whether these two systems represent ends of the spectrum and whether the same TCR might have different energy landscapes depending on the ligand (Fig. 1). One of the ligands key to T cell function is the positive selecting peptide-MHC. Evidence has suggested that germline V regions are evolutionarily predisposed to interact with MHC11. However, V-region CDR1 and CDR2 develop in the presence of CDR3s and thus could evolve not only contacts with MHC but also interactions with neighboring CDR3s. The antigenic selections of specific V-joining (VJ) or V-diversity-J (VDJ) combinations, as shown by Borg et al. and by others, are no doubt a product of the collective interactions and dynamics among the CDRs. In fact, the results with LC13 showing communication between CDR1, CDR2 and CDR3s and a key function for CDR3 in peptide-MHC binding might have been anticipated from another study based on a very different approach12. Restrictions in VJ and VDJ sequences were found at the level of thymic selection and peripheral T cell population expansion12. Predominant CDR3 sequences were often associated with MHC class I or class II restriction (in the context of specific V regions). Several sequences found preferentially in peripheral T cells contained long CDR3s12, perhaps analogous to LC13 and EBV-selected TCRs. It remains to be seen if the populations of T cells that bear the predominant LC13-like TCRs in humans may be preferentially expanded by interactions with a self peptide–HLA-B8 complex.
For interactions with self peptide–MHC during positive selection, the LC13 TCR may or may not adopt a conformation similar to that found in the LC13–FLR-B8 complex (Fig. 1a). If it does, one might predict that the peptide is structurally related to the EBV peptide and that the binding free energies would distribute among CDRs in a way similar to that found for foreign antigen binding (albeit a lower overall affinity). In this case, perhaps Vα residues His33α and His48α are also important in contacting the self peptide (such as a tyrosine at P7). Alternatively, given the structural plasticity observed for TCRs1,2,6, it is possible that a distinct conformation may be adopted after binding self peptide–MHC, and the energy distributions among the CDRs may differ from that of the LC13–FLR-B8 complex. Experiments with self peptide–MHC are not simple, as the positive selecting peptides are typically unknown. Even if they were known, binding affinities of TCRs for self peptide–MHC are so low as to make alanine scans almost impossible. Nevertheless, it will be useful to determine where other TCRs fall in the spectrum of CDR binding energy distributions and to determine how TCRs might vary with different peptide-MHC ligands. Recent advances in computational alanine scanning may help contribute to these analyses7, assuming that the extensive plasticity of TCRs does not prevent reasonably accurate predictions.
TCR differences in the distribution of CDR-derived binding energies could allow reconciliation of various models (two-step binding, V region–biased restrictions and so on). Still, there remains the unanswered issue of the molecular basis of MHC restriction. So far, there have been few conserved TCR contacts observed among TCR–peptide-MHC structures that correlate with MHC restriction2. If germline-encoded CDR1 and CDR2 are only marginally involved in direct binding to MHC helices for some TCRs, then how do these TCRs focus only on MHC and not on any other surface proteins? CD8 and CD4 may be particularly important in these cases, and the energy gained through CDR3-mediated contacts with MHC, together with CD8 or CD4, could drive MHC restriction. Several studies have shown CD8 participates in the binding of peptide-MHC by TCRs on MHC class I–restricted T cells. Modest energies of interaction mediated through CDR3 residues and CD8 could be sufficient to focus T cells on the MHC. Furthermore, thymic selection and peripheral homeostatic expansion of T cell populations most likely require signaling through the recruitment of CD8 or CD4. Binding energy and signaling from the CD8–peptide-MHC interaction may be important in focusing TCRs like LC13 on MHC class I. The findings of Rossjohn, McCluskey and colleagues may refocus attention on the cooperative interactions involving the TCR and these accessory molecules in T cell development.
Rudolph, M.G., Luz, J.G. & Wilson, I.A. Annu. Rev. Biophys. Biomol. Struc. 31, 121–149 (2002).
Bankovich, A.J. & Garcia, K.C. Immunity 18, 7–11 (2003).
Borg, N.A. et al. Nat. Immunol. 6, 171–180 (2005).
Sim, B.-C., Zerva, L., Greene, M.I. & Gascoigne, N.R.J. Science 273, 963–966 (1996).
Wu, L.C., Tuot, D.S., Lyons, D.S., Garcia, K.C. & Davis, M.M. Nature 418, 552–556 (2002).
Reiser, J.B. et al. Nat. Immunol. 4, 241–247 (2003).
DeLano, W.L. Curr. Opin. Struct. Biol. 12, 14–20 (2002).
Manning, T.C. et al. Immunity 8, 413–425 (1998).
Lee, P.U., Churchill, H.R., Daniels, M., Jameson, S.C. & Kranz, D.M. J. Exp. Med. 191, 1355–1364 (2000).
Degano, M. et al. Immunity 12, 251–261 (2000).
Zerrahn, J., Held, W. & Raulet, D.H. Cell 88, 627–636 (1997).
Correia-Neves, M., Waltzinger, C., Mathis, D. & Benoist, C. Immunity 14, 21–32 (2001).
About this article
Nature Communications (2013)